A healthy diet is protective against most major chronic diseases, including obesity, type-II diabetes, and hypertension, and can also benefit mental health, longevity, and overall wellbeing (Centers for Disease Control and Prevention, 2021). However, in the United States and many other countries, few adults meet healthy dietary recommendations and diet-related disease has become a leading cause of death (Anand et al., 2015; Afshin et al., 2019). Although Americans are increasingly consuming foods away from home, prepared by restaurants and fast-food outlets, food prepared at home tend to be more nutritious, less caloric, and more affordable (Saksena et al., 2018). People's capacity to prepare and eat healthy foods depends in part on their access to grocery stores and supermarkets; a primary source of affordable healthy food options in the United States (Glanz et al., 2005; Laska et al., 2010, Walker et al., 2010). Frequent grocery shopping is associated with healthier diets and lower obesity rates (Gustat et al., 2015; He et al., 2012; Minaker et al., 2016; Thornton et al., 2012; Widener et al., 2018). Visiting a wider variety of grocery stores has also been associated with more healthy and balanced diets (Cervigni et al., 2020; Liu et al., 2014; Shearer et al., 2015; Zenk et al., 2011). However, inequities in access to grocery stores are well-documented: People with low incomes and people of color are more likely to live in areas with limited access to grocery stores, which may contribute to disparities in nutrition and diet-related health outcomes among these populations (Bell et al., 2019; Larson et al., 2009; Maguire et al., 2017).

One limitation of existing research on healthy food access has been a focus on people's access to grocery stores near their home. Often studies examine access within home census tracts, ZIP codes or home-centric buffers, implying the assumption that people primarily shop in areas close to where they live (Caspi et al., 2012; Charreire et al., 2010; Feng et al., 2010; Gamba et al., 2015; Leal and Chaix, 2011). However, research measuring grocery store visits has shown that people do not solely shop at stores that are closest proximity to their home (Lucan, 2015; Matthews and Yang, 2013; Browning et al., 2017; Inagami et al., 2006; Widener et al., 2013). One review suggested that static methods focusing on home neighborhoods overestimate the importance of the residential food environment, though the magnitude of this potential bias has yet to be quantified (Cetateanu and Jones, 2016).

To address this measurement issue, the operationalization of the “activity space” concept, which refers to the spatial trajectories of people's daily movements, becomes increasingly valuable in studying human mobility and contextual exposures (Blondel et al., 2015, Matthews and Yang, 2013; Yi et al., 2019; Yi et al., 2024). Over the last two decades, the activity space research has evolved significantly, employing various data sources and methodologies. This includes the use of transportation survey data (e.g., Lee and Kwan, 2011), self-reported household travel surveys (e.g., Browning et al., 2017; Cheng et al., 2020), qualitative interviews (e.g., Hillier et al., 2011), and combination of qualitative methods and geographical information systems to allow participants to manually draw their activity spaces on a map (e.g., Basta et al., 2010; Chaix et al., 2012). Recent advancements in mobility data collection, including GPS-enabled mobile phones (Chang et al., 2022; Gao et al. 2013, 2020; Horn et al., 2023, Xu et al., 2023), wearable location sensors (Kerr et al., 2011; Widener et al., 2018; Yi et al., 2022), and social media check-ins (Nguyen et al., 2017), have paved the way for novel insights into human behaviors, including grocery shopping patterns. These diverse methodologies in activity space research underscore its adaptations to technological advancements in understanding human mobility and spatial behaviors and call for more quantitative research to complement and enhance our understanding of the insights gained from qualitative studies.

Despite these technological advances, the application of mobility data to research has been constrained to limited spatiotemporal scales, primarily due to the challenges associated with data collection, such as the high cost of GPS devices, and the potential for recall bias in data reporting (Alexandre et al., 2020; Browning et al., 2017; Clary et al., 2017; Perchoux et al., 2019; Smith et al., 2019; Zenk et al., 2011). These obstacles underscore the need for innovative approaches to leveraging mobility data more effectively to uncover the complex dynamics of grocery shopping behaviors and their implications on public health and urban planning. Two recent studies have utilized large-scale anonymized and aggregated mobile phone location data, providing evidence that visits to food retailers are a meaningful proxy for dietary intake (Horn et al., 2023), and significantly predict diet-related diseases (Horn et al., 2023; Xu et al., 2023). However, these analyses were limited to fast-food outlet visits. This study aims to apply this promising approach, using large-scale mobility data to offer insights into grocery store visits for large and diverse populations of mobile phone users over long periods of time, rather than short snapshots of behaviors captured by other methods.

Disparities in poor diets and diet-related diseases are pronounced and pervasive, and a lack of access to healthy food is acknowledged as a key “social determinant of health” (Downs et al., 2020; Glanz et al., 2005; Story et al., 2008; Swinburn et al., 2011). As a result, past research has often explored the sociodemographic disparities in grocery shopping behaviors, identifying barriers to grocery shopping, including racial and ethnic minority status (Shier et al., 2022), age (Angell et al., 2012; Netopil et al., 2014; Wu et al., 2022), disability (Charnes 2022), low income (Darko et al., 2013; Zachary et al., 2013), unreliable transportation (Burns et al., 2011; Thompson et al., 2022; Gustat et al., 2015), financial constraints (Inglis et al., 2009; Burns et al., 2011), and low enrollment in food assistance programs (Rose and Richards, 2004; Ver Ploeg et al., 2015). These studies have primarily relied on interviews and surveys to gather insights, focusing largely on individual experiences and perceptions. While valuable, this approach leaves a gap in our understanding that could be addressed through quantitative research, particular concerning how neighborhood characteristics are associated with grocery shopping patterns. These insights are needed to better understand the role of food access as a key social determinant of health that can give rise to disparities in a range of health outcomes and risk for diseases.

To address these gaps, the goal of our study is to use large-scale mobility data, captured over one year (2021), to investigate visits to grocery stores among residents of a historically under-resourced area of Los Angeles County (LAC)'s east side. By linking this mobility data to information about the neighborhood sociodemographic and retail food environments, we will also explore differences and disparities in grocery store use. Specifically, we will address the following two research questions:

What grocery store visit patterns do we observe using mobility data captured over one year?

Are residents' grocery store visit patterns associated with neighborhood sociodemographic and food accessibility?